After a two-day running conversation with my boss, I’ve decided I should probably write about this. Anyway I owe Joe the Bossman Ross credit for this post.
This morning Jamie Redknapp gave us his opinion of Manchester City’s new £34 million acquisition from Shakhtar Donetsk, Fernandinho. Here is a sampling of what Redknapp had to say on the Brazilian midfielder:
“Fernandinho can become a monster. His stand-out Champions League performances for Shakhtar Donetsk suggest he will be a weapon in the Premier League once he comes to terms with the increased fitness demands.”
“Full name Fernando Luiz Rosa (Fernandinho means little Fernando) he is thin up top but his power comes from his thighs and he covers the ground at speed.”
“We pigeon-hole midfielders into what they can and can’t do. This Brazilian can do the lot. He can play one-twos at real pace, drive at opponents, tackle, intercept, win the ball and set up a breakaway. His what City is missing”
“He’s a terrific crosser of the ball…”
You might have noticed what’s not here: any empirical evidence to back up these claims. No attempt at an argument that goes beyond subjective impression based on 8 Champions League appearances (we’re apparently supposed to believe Redknapp has seen all of them). No hard evidence on Fernandinho’s positioning, no numbers on take-ons or cross accuracy, no attempt to make a coherent argument beyond the claim the club ‘over-relies’ on Yaya Toure and that Gareth Barry and Javier Garcia are “too one-paced.”
Redknapp of course isn’t alone in playing Mr. Crystal Ball with a potential star. Every summer there are transfers or rumors of transfers. And because someone decided a long time ago that football fans will apparently drop dead en masse unless a pundit tells them what to think, we get subjected to ‘the experts.’ Hello, I work in media. This player is this. He will offer your club that. I saw them play. You didn’t. I know. Trust me.
For some, that’s good enough. Football’s a romantic sport at its core, unreadable and mysterious, so what are a few vague guarantees of future greatness between friends? It’s not like Redknapp is giving stock market advice. He’s just making a few fans feel better about their club paying a lot of money for a Brazilian who played in the Ukrainian Premier League.
For those looking to make up their own mind, we’re left with a handful of less-than-stellar options to make up our own minds. I’ve taken some time today compiling them for your benefit.
The Media
See Jamie Redknapp. Although some people do indeed know what they’re talking about, tis true. But keep in mind, they have access to exactly the same resources that the rest of us have. So unless they’re an experienced scout or have a good track record, you may as well get your information from some guy.
YouTube Videos
This is the go-to choice for many nervous football fans, but it comes with obvious limitations. Like the fact that a four minute compilation is going to tell you as much about a footballer as a four second test drive will tell you about a car. So the less said about this approach the better.
Their Transfer Market Price
As I pointed out in yesterday’s post, transfer market prices are not static but vary depending on the clubs seeking to sell, the clubs aiming to buy, the relative national tax rate, etc. etc. This doesn’t prevent fans from engaging in endless debate over whether this newcomer from Latvia is really worth Forty million Quatloos.
Or better yet, think of it this way: there is no such thing as an accurate transfer market value based on a player’s objective value as a footballer. If a club pays £50 million for Fernando Torres, that’s exactly what he’s worth. Why? Because that’s what the club paid for him. It’s tautologically delicious!
If a club bought a player, then they must have some hidden proprietary value that we don’t know of. Trust them to make the right decision.
Two words for you my naive reader: Damien. Comolli. The former Liverpool sporting director famously (and perhaps apocryphally) purchased players based on their average final third entries. Alone. Some clubs have good scouts, some don’t. Some make their decisions based both on experience, psychology, and some reliable underlying metrics, and even then they make mistakes. But most fans know this of course.
Besides, if you’re even engaging in this kind of debate, it means you believe in something called “true value”, which means the transfer fee itself is moot.
Player Data
Right now there are several websites with reams of data that can tell us everything about a footballer, except for why. From Whoscored, we can tell who won the most aeriel duels. On Squawka, we can see if a player took more shots on goal with their left peg or their right.
Publicly available football data is never a bad thing, but the problem right now is that we still have no clue what makes for good data and what doesn’t. Reading through Whoscored’s metrics, a fan might look at one measurement in isolation—say their average number of tackles per game(if we’re looking at a defender)—notice that the number is high, and then use this number to make the claim: “this is a good defender.”
I’ve never coined an expression, so this might be a unique first: let’s call this the Jaap Stam Fallacy. Stam was the Manchester United defender who was famously axed by Sir Alex Ferguson for not tackling enough. Simon Kuper fills us in:
In August 2001 Manchester United’s manager Alex Ferguson suddenly sold his defender Jaap Stam to Lazio Roma. The move surprised everyone. Some thought Ferguson was punishing the Dutchman for a silly autobiography he had just published. In truth, although Ferguson didn’t say this publicly, the sale was prompted partly by match data. Studying the numbers, Ferguson had spotted that Stam was tackling less often than before. He presumed the defender, then 29, was declining. So he sold him.
Of course Stam wasn’t a bad defender; the lack of tackles came from discipline in positioning. He didn’t tackle because he didn’t need to tackle. The lesson here is football is a complex sport. Everything is connected. This doesn’t mean we can’t isolate individual metrics and use them to draw a conclusion about a player’s worth. The problem is however that we don’t know with a great deal of certainty which ones to isolate with any meaning, and under what context. That’s before we get into a host of other statistical problems; like for example how actions-per-game are calculated (it should be per 90 minutes, not per-game, to account for the imbalance of stoppage time).
Additionally, there’s the issue of players coming in from leagues of which we have precious little data. Fernandinho is a good example, seeing as he’s arriving to City from Shakhtar. There are few published data on shots, key passes, or other popular player metrics on any data website I’ve yet seen on the Ukrainian Premier League. Or indeed many of the leagues from which stars are increasingly culled. You’d think it would make more sense to have data available on leagues that rarely enjoy TV deals, but that’s a topic for another day.
Your Own Eyes
This is the option of choice for most sane people, although it still comes with all the same issues as the above. You’ve watched Jesus Navas play for a long time, you know exactly how capable he is as a player, you know how City plays, you have a reasonably good idea of how Pellegrini might use him. You feel good about yourself. Have a gold star sticker!
However this works best if a) you’ve watched a lot of Sevilla lately, b) you’ve watched a lot of City lately, c) you’ve watched Pellegrini do his thing week-in and week-out at Malaga, and d) you know a lot about what teams should generally do to win at the football. A tiny few of us can claim all four categories. Some of us have boned up on a few of these areas but not others. Still more of us are kind of okay in all of them.
In the end most of us have enough knowledge here to fuel a debate with our friends. And that’s really all football should be for. It’s a game. But shouldn’t we hope for something more than this? Probably not, but anyway…
Football Analytics
Where Data Becomes Knowledge™. This is probably the last hope for a Unified Theory of Footballer Goodness, but we should be careful never, EVER to reduce players to a set of isolated metrics. Still, it’s good to have something to bridge the speculative and subjective with the empirical and the concrete. Soccer analytics is nowhere near doing that well-enough just yet, but there have been some interesting movements in that direction lately.
I’ve already mentioned Ted Knutson’s work looking at key passes and assists in evaluating attacking midfielders.
Even today the blog 11tegen11 looked at the mean conversion rate of shots by area of the pitch, kind of crude X,Y positioning chart. The blogger’s conclusion?
These strike zones will prove a useful reference point to assess difference between teams in terms of strategy, difference between players in terms of quality and will help to explore our beautiful game more in depth.
Wah-hey! Now we’re getting somewhere. Although nowhere near close to an ideal. But still, there is some hope that we won’t be eternally left groping in the dark or stuck drooling over looped b-reels of the same highlights again.
This morning Jamie Redknapp gave us his opinion of Manchester City’s new £34 million acquisition from Shakhtar Donetsk, Fernandinho. Here is a sampling of what Redknapp had to say on the Brazilian midfielder:
“Fernandinho can become a monster. His stand-out Champions League performances for Shakhtar Donetsk suggest he will be a weapon in the Premier League once he comes to terms with the increased fitness demands.”
“Full name Fernando Luiz Rosa (Fernandinho means little Fernando) he is thin up top but his power comes from his thighs and he covers the ground at speed.”
“We pigeon-hole midfielders into what they can and can’t do. This Brazilian can do the lot. He can play one-twos at real pace, drive at opponents, tackle, intercept, win the ball and set up a breakaway. His what City is missing”
“He’s a terrific crosser of the ball…”
You might have noticed what’s not here: any empirical evidence to back up these claims. No attempt at an argument that goes beyond subjective impression based on 8 Champions League appearances (we’re apparently supposed to believe Redknapp has seen all of them). No hard evidence on Fernandinho’s positioning, no numbers on take-ons or cross accuracy, no attempt to make a coherent argument beyond the claim the club ‘over-relies’ on Yaya Toure and that Gareth Barry and Javier Garcia are “too one-paced.”
Redknapp of course isn’t alone in playing Mr. Crystal Ball with a potential star. Every summer there are transfers or rumors of transfers. And because someone decided a long time ago that football fans will apparently drop dead en masse unless a pundit tells them what to think, we get subjected to ‘the experts.’ Hello, I work in media. This player is this. He will offer your club that. I saw them play. You didn’t. I know. Trust me.
For some, that’s good enough. Football’s a romantic sport at its core, unreadable and mysterious, so what are a few vague guarantees of future greatness between friends? It’s not like Redknapp is giving stock market advice. He’s just making a few fans feel better about their club paying a lot of money for a Brazilian who played in the Ukrainian Premier League.
For those looking to make up their own mind, we’re left with a handful of less-than-stellar options to make up our own minds. I’ve taken some time today compiling them for your benefit.
The Media
See Jamie Redknapp. Although some people do indeed know what they’re talking about, tis true. But keep in mind, they have access to exactly the same resources that the rest of us have. So unless they’re an experienced scout or have a good track record, you may as well get your information from some guy.
YouTube Videos
This is the go-to choice for many nervous football fans, but it comes with obvious limitations. Like the fact that a four minute compilation is going to tell you as much about a footballer as a four second test drive will tell you about a car. So the less said about this approach the better.
Their Transfer Market Price
As I pointed out in yesterday’s post, transfer market prices are not static but vary depending on the clubs seeking to sell, the clubs aiming to buy, the relative national tax rate, etc. etc. This doesn’t prevent fans from engaging in endless debate over whether this newcomer from Latvia is really worth Forty million Quatloos.
Or better yet, think of it this way: there is no such thing as an accurate transfer market value based on a player’s objective value as a footballer. If a club pays £50 million for Fernando Torres, that’s exactly what he’s worth. Why? Because that’s what the club paid for him. It’s tautologically delicious!
If a club bought a player, then they must have some hidden proprietary value that we don’t know of. Trust them to make the right decision.
Two words for you my naive reader: Damien. Comolli. The former Liverpool sporting director famously (and perhaps apocryphally) purchased players based on their average final third entries. Alone. Some clubs have good scouts, some don’t. Some make their decisions based both on experience, psychology, and some reliable underlying metrics, and even then they make mistakes. But most fans know this of course.
Besides, if you’re even engaging in this kind of debate, it means you believe in something called “true value”, which means the transfer fee itself is moot.
Player Data
Right now there are several websites with reams of data that can tell us everything about a footballer, except for why. From Whoscored, we can tell who won the most aeriel duels. On Squawka, we can see if a player took more shots on goal with their left peg or their right.
Publicly available football data is never a bad thing, but the problem right now is that we still have no clue what makes for good data and what doesn’t. Reading through Whoscored’s metrics, a fan might look at one measurement in isolation—say their average number of tackles per game(if we’re looking at a defender)—notice that the number is high, and then use this number to make the claim: “this is a good defender.”
I’ve never coined an expression, so this might be a unique first: let’s call this the Jaap Stam Fallacy. Stam was the Manchester United defender who was famously axed by Sir Alex Ferguson for not tackling enough. Simon Kuper fills us in:
In August 2001 Manchester United’s manager Alex Ferguson suddenly sold his defender Jaap Stam to Lazio Roma. The move surprised everyone. Some thought Ferguson was punishing the Dutchman for a silly autobiography he had just published. In truth, although Ferguson didn’t say this publicly, the sale was prompted partly by match data. Studying the numbers, Ferguson had spotted that Stam was tackling less often than before. He presumed the defender, then 29, was declining. So he sold him.
Of course Stam wasn’t a bad defender; the lack of tackles came from discipline in positioning. He didn’t tackle because he didn’t need to tackle. The lesson here is football is a complex sport. Everything is connected. This doesn’t mean we can’t isolate individual metrics and use them to draw a conclusion about a player’s worth. The problem is however that we don’t know with a great deal of certainty which ones to isolate with any meaning, and under what context. That’s before we get into a host of other statistical problems; like for example how actions-per-game are calculated (it should be per 90 minutes, not per-game, to account for the imbalance of stoppage time).
Additionally, there’s the issue of players coming in from leagues of which we have precious little data. Fernandinho is a good example, seeing as he’s arriving to City from Shakhtar. There are few published data on shots, key passes, or other popular player metrics on any data website I’ve yet seen on the Ukrainian Premier League. Or indeed many of the leagues from which stars are increasingly culled. You’d think it would make more sense to have data available on leagues that rarely enjoy TV deals, but that’s a topic for another day.
Your Own Eyes
This is the option of choice for most sane people, although it still comes with all the same issues as the above. You’ve watched Jesus Navas play for a long time, you know exactly how capable he is as a player, you know how City plays, you have a reasonably good idea of how Pellegrini might use him. You feel good about yourself. Have a gold star sticker!
However this works best if a) you’ve watched a lot of Sevilla lately, b) you’ve watched a lot of City lately, c) you’ve watched Pellegrini do his thing week-in and week-out at Malaga, and d) you know a lot about what teams should generally do to win at the football. A tiny few of us can claim all four categories. Some of us have boned up on a few of these areas but not others. Still more of us are kind of okay in all of them.
In the end most of us have enough knowledge here to fuel a debate with our friends. And that’s really all football should be for. It’s a game. But shouldn’t we hope for something more than this? Probably not, but anyway…
Football Analytics
Where Data Becomes Knowledge™. This is probably the last hope for a Unified Theory of Footballer Goodness, but we should be careful never, EVER to reduce players to a set of isolated metrics. Still, it’s good to have something to bridge the speculative and subjective with the empirical and the concrete. Soccer analytics is nowhere near doing that well-enough just yet, but there have been some interesting movements in that direction lately.
I’ve already mentioned Ted Knutson’s work looking at key passes and assists in evaluating attacking midfielders.
Even today the blog 11tegen11 looked at the mean conversion rate of shots by area of the pitch, kind of crude X,Y positioning chart. The blogger’s conclusion?
These strike zones will prove a useful reference point to assess difference between teams in terms of strategy, difference between players in terms of quality and will help to explore our beautiful game more in depth.
Wah-hey! Now we’re getting somewhere. Although nowhere near close to an ideal. But still, there is some hope that we won’t be eternally left groping in the dark or stuck drooling over looped b-reels of the same highlights again.
